|Health Economics for Developing Countries: A Survival Kit (London School of Hygiene and Tropical Medicine, 1998, 134 p.)|
The reasons for studying resources and costs have been explained in Chapter 6. It is also worth noting that most health sectors consume between 3 and 10% of GNP. This represents an enormous resource commitment each year which, under the combined pressures of economic, technological and demographic changes, has, until recently, been rising rapidly in many countries.
The usual reasons for studying costs, such as the need to know the resources necessary for existing and future activities and the need to monitor costs in order to achieve efficiency have, therefore, recently been supplemented by the growing need to control resource use and costs.
This chapter provides a brief overview of the economist's view of costs, how they are used and how they are calculated.
This term refers to the various inputs or factors of production which comprise the raw ingredients of any productive activity. A traditional distinction is drawn between three types of factor: land, labour and capital.
Interest lies not so much in the resources themselves, but in their potential contribution to the production of socially valued outputs (such as goods and services) which satisfy human wants and desires. For all intents and purposes the amount of resources available to satisfy these wants is finite, whereas many of the wants are infinite. It is this problem of scarcity that requires consideration of the costs associated with the use of resources for a particular purpose.
Economists are concerned, therefore, with the way that productive activities can be undertaken with different combinations of resource inputs. Typical problems studied are:
- the effect on output of employing different proportions of human and capital inputs
- the possibility for increasing output through specialization and the division of labour
- the effect on output of increasing the input of some resources while holding other inputs constant.
Most people are familiar with the concept of cost as the sacrifice necessary in order to obtain a good or a service. This is also the way that economists understand the term. The difference between economists and others lies in their understanding of the nature of the concept of sacrifice.
In general, people encounter costs in terms of the monetary prices placed upon goods and services, and so they equate sacrifice with price. This is a valid if rather narrow approach. Economists argue that the sacrifice to be considered is the alternative opportunities that are given up when a choice is made to use a resource or resources in a particular way. Choice is exclusive by nature and the opportunity cost represents the value of the sacrifice involved in it. For example, if a budget of say, US$1m could be used in order to build and run either a rural health service or a car plant, then the economic (opportunity) cost of the health service is the car plant that has to be given up.
In some cases the money price of a good or service may be a good indicator of this opportunity cost, but there are likely to be instances where it is not. For example:
- in certain situations resources will be used that do not have a money price. Goods and services given freely do, however, have alternative uses to which they can be put, and it is the value of these alternative uses that is the true measure of sacrifice or costs
- certain resources do have a monetary price but, for various reasons, this price does not accurately reflect the value of their alternative productive uses
- certain activities may have 'spillover-effects' on other producers or consumers. These effects, called externalities, may be positive (ie. beneficial), or negative (ie. harmful). Money prices may reflect only the private costs (or benefits) for producers and consumers, and not allow for the possible existence of social costs (or benefits)
- money prices are often distorted by transfer payments, the redistribution of income from one group to another. Transfers such as taxation or subsidies do not in themselves constitute an opportunity cost and so should be excluded from its calculation.
4. Cost Concepts and Cost Analysis
Cost analysis can focus on many different issues, two of which are explored here to explain cost concepts and to illustrate the measures used in cost analysis. These issues are the relationship between cost and output and the between cost and the time horizon.
4.1 Cost-output relationships
In studying the relationship that exists between cost and output it is useful to separate the analysis into two component parts. The first looks at the physical relationship that exists between inputs and outputs or what is termed the production function. The second looks at the costs attached to those inputs.
Consider the physical production function first. In its simplest form, production consists of using two inputs to produce some good or service. As one input varies, three alternatives are possible within the quantitative relationship between inputs and outputs. Output can rise proportionately faster than the input, proportionately slower, or at the same rate.
This relationship has been much studied within economics. It is the subject of a famous economic hypothesis (the hypothesis of diminishing marginal returns) which postulates that:
If increasing amounts of a variable factor are applied to a fixed quantity of another factor, the output per unit of the variable factor will eventually decrease.
Empirical evidence shows that this hypothesis holds true in a wide range of circumstances. The hypothesis, however, only suggests that output per unit of the variable factor will eventually fall. At any particular point, total output may be rising due to the effects of specialization or the division of labour.
The second component part of the relationship between output and cost is the relationship between the resource inputs and price. Although it is frequently assumed that the price of resources will remain constant as output rises, it is possible that they will not. They may fall if it is possible to negotiate discounts through bulk purchases; or they may rise if more needs to be paid to attract scarce resources as the demand for them increases.
The overall relationship between output and cost is determined by the interaction of these two components. The cost of a particular level of output is the result of the quantities of resources used to create it multiplied by their price. This overall cost will vary as either the productivity of the inputs varies, or as their price varies, or as both vary.
A variety of problems are typically studied by economists: How do costs vary with the level of output? What level of resources is required to achieve different levels of output? How do the costs of each unit of output vary with the scale of activity? How many resources are associated with a small incremental change in the level of activity? In order to facilitate such analysis, three measures of cost are useful:
- total cost (TC) is a measure of all the costs entailed in producing a given level of output. It is a measure of the aggregate resource requirements of a particular scale of activity and is derived by summing all the costs incurred during production. An example would be the cost of providing an immunization service.
- average cost (AC) is a measure of the total costs of production associated with each unit of output. Average costs indicate the resource requirements for each unit of output and are calculated by dividing total costs (TC) by the number of units of output (Q) (AC = TC/Q). An example would be the cost per immunization provided.
- marginal cost (MC) is a measure of the resources associated with a small incremental change in output. Most economic decisions are not about whether to produce all or nothing, but concern small changes in the existing scale of activity. Marginal cost is a measure of the change in costs associated with increasing or decreasing output by one, and is derived by calculating the change in total costs for that one unit. In effect, the marginal cost concept is an attempt to measure the rate of change of costs as output changes. Without a large number of observations and the use of statistical techniques it is difficult to calculate the change in costs for one unit. Therefore, the change in costs for several units is often calculated, and a measure of the incremental unit cost of that larger variation (obtained by dividing incremental costs by incremental units of output) is used instead of marginal cost. An example would be the incremental cost of adding an additional vaccine to the existing service, or of extending the service to an additional village.
The relationship between these three measures can cause some confusion, although in practice it is quite straightforward. If output is increased, say by one unit, total costs will rise by the marginal cost of that unit. How will this affect average costs? If the marginal cost of this extra unit is the same as the previous average cost then the new average cost will be unchanged. However, if the marginal cost of the extra unit is smaller (larger) than the average cost, the new average cost will fall (rise).
These relationships between the different cost measurements can be visualized best in average and marginal cost curves which show how average and marginal costs change as output changes (Figure 17). The average cost curve is hypothesized to be U-shaped, being dominated at low outputs by the effect of the spreading of fixed costs, and at high outputs by the effect of diminishing returns. The marginal cost curve passes as a matter of mathematical necessity through the minimum point of the AC curve, with MC less than AC at lower outputs and greater than AC at higher outputs.
4.2 Costs and the time horizon - fixed and variable costs
One of the purposes of studying costs is to assist decision-making and so it is important to identify those areas over which choice can be exercised. In practice, some resources will have been committed to particular activities by past decisions, and likewise, present decisions will have resource implications well into the future. It is often useful, therefore, to draw a distinction between those resources that are easily varied, and those that are relatively fixed. This classification is crucially dependent upon the time period that is adopted, because in the very long-run, most things can be varied.
Although in practice decisions will be made over a time continuum, economists find it useful to distinguish between a number of discrete decision periods. The two most commonly used periods are the short-run and the long-run.
In the short-run it is assumed that certain resources are fixed and cannot be transferred from their present use. The only way that output can be varied during this period is by changing the amount of variable factors. This is different from the long-run decision period in which all factors are assumed to be variable. The actual time period corresponding to these periods will vary according to how quickly the quantity of fixed factors can be changed. This 'fixedness' can be demonstrated to have important consequences for the output/cost relationship.
In the short-run, with a given level of fixed resource inputs, there is a level of variable resources for which cost per unit of output produced is minimized. Because some inputs are fixed, variations in short-run output can only be achieved by changing the variable resources and thereby altering the proportions of those factors. In doing so there is a move away from the short-run optimum. Short-run variations in output can only be achieved at the expense of increases in unit costs (i.e. the law of diminishing returns).
In the long-run, however, all inputs can be varied. For each level of variable inputs it is also possible to use a level of fixed resources which represents the optimal combination necessary to achieve the lowest unit cost for that level of output. Since in the long-run and the short-run there will be points at which the level of fixed resource inputs coincide, the long-run cost curve can be thought of as the sum of optimal points for all short-run cost curves.
It is possible to study the relationship between input and output as the scale of production varies. When inputs are increased in the same proportions it is possible for output to increase at the same rate (constant returns to scale), at a lower rate (diseconomies of scale), or at a higher rate (economies of scale). Factors such as specialization and the division of labour may account for increasing returns at lower levels of activity. Thereafter, as scale increases, managerial inefficiency and the problems of large scale production are seen as eventually leading to diseconomies of scale.
5. Components of Cost
Some terms have been used above which it is important to define clearly.
5.1 Fixed and variable costs
Fixed costs refer to the costs that are necessarily incurred in setting up a productive activity. Often known as 'overheads', fixed costs are born irrespective of how much output is produced. Variable costs, conversely, are those costs that are directly related to how much output is produced. As most factors of production can be moved to another activity in the long term, fixed costs are usually defined in relation to the time period under consideration (e.g. the financial year).
5.2 Capital and recurrent costs
Capital costs are usually defined as costs for items with a life of more than one year. Typical examples include costs for the:
- construction of buildings
- purchase of equipment
- basic manpower training.
Recurrent costs are the costs that are necessarily incurred each year. These include items such as:
- salaries and wages
- supplies (e.g. drugs, petrol)
- electricity, water
- in-service training.
Most capital items will fall into the category of fixed costs.
6. Comparing Costs
In order to compare costs between institutions, programmes or countries, various adjustments must be made to cost measures. Two adjustments considered here are making capital costs comparable with recurrent costs, and enabling comparisons to be made over time and between countries.
6.1 Comparing capital costs with recurrent costs
The distinction between capital and recurrent costs is important. Capital costs are typically borne in the year the items are acquired and yet those items will provide a service after that time. In order to provide a comprehensive picture of resource use (as opposed to the picture of the timing of expenditures required for budgets), capital costs need to be made comparable with recurrent costs. This can be done by translating the cost of land, buildings and equipment into an equivalent annual cost that takes into account both the expected life of the capital item and a rate of interest (to reflect the opportunity cost of money used for capital purchases that could be used for other purposes).
The simplest way of calculating the equivalent cost of a capital item that yields services over more than one year is to use published tables of present worth factors which provide ready-made calculations for different rates of interest. The tables consist of a matrix with years of life on one axis and interest rates on the other. The initial capital cost is divided by the present worth factor for the number of years the item will be used and the chosen interest rate.
If the cost of tying up funds in capital goods is ignored then the method of straight-line depreciation could be employed to obtain an annual value for capital cost. This simply involves dividing the initial cost of the item by the number of years of expected life.
6.2 Longitudinal and cross-sectional cost comparisons
In many instances it is useful to make a comparison of costs at two different times, or even between two different countries. This is difficult for a number of reasons. Comparing costs over time has to allow for the effects of inflation and changes in relative prices; whereas comparisons across countries have to adjust for different currencies. In order to understand better the issues involved in such comparisons it is useful to repeat the economist's understanding of cost.
Used by economists, the term 'cost' refers to the sacrifice made in order to obtain something else; that is, its opportunity cost. This is an attempt to measure the 'real' rather than the 'nominal' (or monetary) cost attached to particular activities. In comparing costs over time, inflation disguises the real changes in costs and purchasing power by changing the real value of money. It is possible to compensate for these inflationary changes through the use of price indices or deflators. These attempt to reveal true changes in costs over time by controlling for changes in overall or relative prices, or both. Most governments calculate a number of price indices which can be used to determine changes in real resource use rather than merely changes in price.
For example, the Gross Domestic Product Deflator provides an index of the change in price of the resources produced in an economy each year. A more familiar price index is the Retail Price Index which identifies (for a representative sample of goods and services consumed by a household) the annual changes in nominal prices, and indirectly, the real resources that it is possible to obtain with a given budget.
A persistent problem in calculating changes in real resources within the health sector is that of devising a price index that is representative of the diverse range of goods and services that it purchases.
Similar considerations apply to making comparisons of costs across geographical regions. One problem is that the currency differs, making direct comparison of resource use impossible. Exchange rates may provide a highly imperfect means of comparison because of under-/over-valuation of the domestic currency.
A related problem is that the internal price structure of countries may differ in such a way as to change the relative values of different resources. This particularly affects the price of inputs such as manpower which are not easily traded internationally. For example, a given expenditure in dollar terms on physicians' services may purchase a very different quantity of services in different countries, whereas a given expenditure on drugs may purchase similar quantities.
Even determining the overall level of resources available across countries has proved difficult. In developing countries a high proportion of economic activity may not be adequately reflected in figures for national income or expenditure because it occurs within the informal sector of the economy (e.g. subsistence farming). This will distort the calculation of real resource availability and render international comparison difficult.
Economists have attempted to overcome these difficulties surrounding the valuation of resources and international comparisons by calculating purchasing power parities which reflect the real resources available to countries and to sectors - such as the health sector - within them. Considerable conceptual and methodological problems exist in developing and using these however, and problems of currency conversion remain a persistent problem in international cost comparisons (see Chapter 9 for other approaches to obtaining 'true' measures of opportunity cost).
7. Costing Methods
There are two general approaches to collecting and using cost data in the health sector.
7.1 The direct accounting approach
This approach focuses on the costs directly associated with a particular activity. It requires the identification and measurement of the costs incurred in providing a particular intervention or in treating a particular patient. In other words, it attempts to calculate the costs of the resources that are actually used in an activity or those that are consumed by a particular patient. In practice this often proves difficult because it is difficult to identify the resources used in many activities and to calculate what proportion of the costs of shared items or facilities should be apportioned to individual activities.
Typical examples of the direct accounting approach include:
- activity costing - detailed costing of the activities of different cost centres, such as whole hospitals, speciality departments, or health centres
- disease costing - attempts to identify the cost incurred in treating particular diseases, or particular types of patient, or even in carrying out certain types of procedure.
The direct accounting approach is usually the basis for calculating costs in economic evaluations (e.g. cost-effectiveness and cost-benefit studies).
7.2 The statistical approach
This approach focuses on the costs associated with types of activity or types of patient and not on the costs associated with individual activities or individual patients. Rather than addressing the issue of how much something costs and how these costs are incurred, the statistical approach addresses the issues of why costs differ, and by how much.
The technique used within the statistical approach is to estimate the qualitative and quantitative impact on costs of differences between health units or facilities. Multiple regression analysis is frequently used in order to compare cost profiles across a large range of different units. The aim is to eliminate the possible effects on costs of all factors except the one under study. This allows a simulated comparison of like with like, where it would otherwise be impossible. The remaining cost difference between health units/facilities is arguably a measure of the cost implication of the variable under study.
The statistical approach can be illustrated by the following example. In studying, say, the costs associated with providing hospital care, it is very likely that there will be wide variations in cost between different hospitals. Why should this be the case? Some hospitals may be more effective or more efficient than others, or there may be other reasons why these costs vary. Hospitals may differ with respect to:
- size (number of beds)
- throughput (the number of cases treated in a given time period)
- case-mix (the combination of diagnoses admitted)
- case-severity (more/less severe cases admitted)
- quality of care;
- type of treatments offered
- teaching or research activities
- age and/or location of hospital facilities
- occupancy levels
- manpower availability
- length of stay.
One or more of these factors may account for differences in measured costs between hospitals. In studying these cost variations it is necessary to ask what factors are likely to influence costs, and then try to isolate the effect of those factors using statistical techniques such as multiple regression analysis. For instance, it might be interesting to know whether the cost per patient differs according to whether the hospital is large or small. By controlling in the analysis for other factors that affect cost, the effect on cost of size alone can be examined.